# Data dictionary

Every value cell is preserved exactly from the extraction run. `source_document` is the original browser PDF file name.

| Column | Meaning |
| --- | --- |
| `source_document` | Original source PDF file name for the row. |
| `source_pages` | Number of extracted source pages. |
| `Company Name` | First trimmed value from a case-insensitive literal `Company Name : value` line or same-line column-gap pair; blank when not found. |
| `Company Name_input_page` | Internal extraction page used to derive evidence. |
| `Company Name_input_literal_match` | `true` only for an exact literal field-line match. |
| `Company Name_page` | One-indexed source page where the value was read; `0` when not found. |
| `Company Name_evidence` | `cited`, `uncited`, or `not_found` according to the literal-match rule. |
| `Contract Date` | First trimmed value from a case-insensitive literal `Contract Date : value` line or same-line column-gap pair; blank when not found. |
| `Contract Date_input_page` | Internal extraction page used to derive evidence. |
| `Contract Date_input_literal_match` | `true` only for an exact literal field-line match. |
| `Contract Date_page` | One-indexed source page where the value was read; `0` when not found. |
| `Contract Date_evidence` | `cited`, `uncited`, or `not_found` according to the literal-match rule. |
| `Contract Value` | First trimmed value from a case-insensitive literal `Contract Value : value` line or same-line column-gap pair; blank when not found. |
| `Contract Value_input_page` | Internal extraction page used to derive evidence. |
| `Contract Value_input_literal_match` | `true` only for an exact literal field-line match. |
| `Contract Value_page` | One-indexed source page where the value was read; `0` when not found. |
| `Contract Value_evidence` | `cited`, `uncited`, or `not_found` according to the literal-match rule. |
| `Signatory Name` | First trimmed value from a case-insensitive literal `Signatory Name : value` line or same-line column-gap pair; blank when not found. |
| `Signatory Name_input_page` | Internal extraction page used to derive evidence. |
| `Signatory Name_input_literal_match` | `true` only for an exact literal field-line match. |
| `Signatory Name_page` | One-indexed source page where the value was read; `0` when not found. |
| `Signatory Name_evidence` | `cited`, `uncited`, or `not_found` according to the literal-match rule. |
| `Governing Law` | First trimmed value from a case-insensitive literal `Governing Law : value` line or same-line column-gap pair; blank when not found. |
| `Governing Law_input_page` | Internal extraction page used to derive evidence. |
| `Governing Law_input_literal_match` | `true` only for an exact literal field-line match. |
| `Governing Law_page` | One-indexed source page where the value was read; `0` when not found. |
| `Governing Law_evidence` | `cited`, `uncited`, or `not_found` according to the literal-match rule. |

## Literal-match normalization rule

Field matching is case-insensitive and regexp-escapes the requested field text. The plain form accepts a complete field line with optional whitespace around the colon and a non-empty value; layout-preserved text also accepts a single field/value pair separated by a two-or-more-space column gap. Prose, headings, and near matches do not qualify.